The discovery that samples of many types of human tumor will grow progressively when implanted into athymic 'nude' mice represents a milestone in the study of human neoplasia. Though this technique - referred to as 'xenografting' - is used throughout the world as a routine method in cancer research, it nevertheless suffers from one serious limitation, and that is a failure of the implanted tumor to metastasize in almost all instances. Thus, such grafts assume a non-malignant behaviour even if the tumor was highly malignant in its original (human) host. When metastases do manifest themselves, the pattern is usually not reflective of the clinical behaviour of the tumor. Our aim is to isolate and characterize heritable and phenotypically stable mutant cell lines from human melanomas (and eventually from other types of tumors) which not only will readily metastasize in nuce mice, but which will do so in a manner that faithfully reproduces their usual clinical behaviour. Our guiding hypothesis, based primarily on recent work in our laboratory, is that such cell subpopulations exist in very low frequencies but can be isolated through the selection of so-called lectin resistant (Lec-r) 'glycosylation mutants'. These are cells which usually possess cell surface carbohydrate alterations the nature of which, at least in some cases, can endow cells with unique metastatic abilities, (including tumor cells of human origin growing in nude mice). Because Lec-r mutants appear to arise in most instances as a consequence of point or deletion mutations, the cell populations comprising them are invariably stable and highly homogeneous for their altered - and sometimes unique - lectin resistant and metastatic phenotypes. Thus, they are excellent cell populations with which to initiate studies aimed at identifying cell surface biochemical (structural) features and genes relevant to the expression of the metastatic phenotype- the second major aim of this proposal. Clearly, the availability of such well-defined human tumor mutant cell populations displaying clinically-relevant metastatic phenotype in nude mice would be a valuable resource for human cancer cell biology research studies, in general and for oncology studies involving experimental therapeutics in vivo, in particular.